In recent information processing systems, a plurality of types of apparatuses (for example, storage apparatus, switch, server computer, and so on) are connected to a network and operated. As the system scale increases, the number of apparatuses operated also increases, which makes operational management burdensome. Thus, a management apparatus that supports operational management of such an information processing system is used.
For example, there has been proposed a network manager that is connected to a storage network including a storage system, a server, and a switch. The network manager includes event dictionaries corresponding to products of multiple vendors, and uses the event dictionaries to interpret an event message received from a device experiencing a failure.
There has also been proposed a storage area network (SAN) manager that interprets a failure notification message received from each apparatus on a SAN. The SAN manager detects the influence of a failure on a real volume or a virtual volume.
There has been also proposed a management server that receives information transmitted from an agent in each host computer and a property information notification unit in each hardware apparatus. The management server centrally manages pieces of information transmitted from a plurality of apparatuses by associating the pieces of information with each other in a unique data structure.
See, for example, Japanese Laid-open Patent Publications No. 2004-133897, No. 2007-172003, and No. 2007-257645.
An information processing apparatus includes a plurality of components such as a processing device, a storage device, a communication interface, and so on. The states of these components may change in response to a failure or the like. One method to deal with this may be to cause a management apparatus to periodically collect information on all the components existing in a system, and register and centrally manage the collected information as management information. However, in normal time or the like, for example, there may be no difference between information to be collected and the previously collected information. That is, the management apparatus might perform unneeded information collection.
Another method may be, for example, to cause a management apparatus to notify the system administrator of occurrence of an error, and collect information on all the components in response to an input of a collection instruction from the administrator. However, with this method, although information collection is performed only on limited occasions, a determination of whether to collect information is made by the administrator. Therefore, management information is not updated unless the administrator issues an instruction, which might reduce the accuracy of management information.